Clinical and pathological characterization of Central Nervous System cryptococcosis in an experimental mouse model of stereotaxic intracerebral infection

Infection of the Central Nervous System (CNS) by the encapsulated fungus Cryptococcus neoformans can lead to high mortality meningitis, most commonly in immunocompromised patients. While the mechanisms by which the fungus crosses the blood-brain barrier to initiate infection in the CNS are well recognized, there are still substantial unanswered questions about the disease progression once the fungus is established in the brain. C. neoformans is characterized by a glucuronoxylomannan (GXM)-rich polysaccharide capsule which has been implicated in immune evasion, but its role during the host CNS infection needs further elucidation. Therefore, the present study aims to examine these key questions about the mechanisms underlying cryptococcal meningitis progression and the impact of fungal GXM release by using an intracerebral rodent infection model via stereotaxic surgery. After developing brain infection, we analyzed distinct brain regions and found that while fungal load and brain weight were comparable one-week post-infection, there were region-specific histopathological (with and without brain parenchyma involvement) and disease manifestations. Moreover, we also observed a region-specific correlation between GXM accumulation and glial cell recruitment. Furthermore, mortality was associated with the presence of subarachnoid hemorrhaging and GXM deposition in the meningeal blood vessels and meninges in all regions infected. Our results show that using the present infection model can facilitate clinical and neuropathological observations during the progression of neurocryptococcosis. Importantly, this mouse model can be used to further investigate disease progression as it develops in humans.


Introduction
Cryptococcus neoformans is an encapsulated opportunistic yeast-like fungus that triggers lifethreatening meningoencephalitis in both immunocompromised and healthy individuals. Of the~220,000 annual cases of cryptococcal meningitis in HIV + /AIDS victims worldwide, 181,000 (~82%) resulted in death [1]. Cryptococcus species are the third most common invasive fungus in solid organ transplantation (SOT) recipients [2]. The prevalence of cryptococcosis in this population ranges between 0.2 to 5.8%, with a total mortality rate of 20% to 50%, highlighting its prevalence in another patient population at risk [2,3]. C. neoformans infects humans when desiccated or poorly encapsulated yeasts/basidiospores are inhaled [4,5]. Most immunocompetent hosts render C. neoformans dormant in their lungs, but immunosuppression can trigger its replication and dissemination via blood or lymph to other organs, especially the brain. Approximately 50%-75% of SOT recipients with cryptococcosis have extrapulmonary or disseminated disease [6,7]. During cryptococcosis, fungemia is detected in~50% of HIV-infected patients [8], correlates experimentally with dissemination and brain invasion [9], and is an independent parameter of early mycological failure in humans [8]. C. neoformans crosses the blood-brain barrier [10] via multiple mechanisms: transcytosis, paracellular transit, or as "Trojan horse" cargo within host inflammatory cells [11][12][13][14]. Once in the brain, C. neoformans triggers meningoencephalitis, the most serious hallmark of cryptococcosis.
With a protective polysaccharide capsule, C. neoformans grows at mammalian physiological temperature and produces melanin to aid in virulence [15]. The capsule is the most distinctive physical structure of the cryptococcal cell and contributes substantially to its virulence [16][17][18]. Compared to in vitro conditions, most C. neoformans strains form larger capsules in tissue [17,19], where capsule size can be influenced by physiological CO 2 levels [20], serum [21], osmolarity [22,23], and iron levels [24]. Glucuronoxylomannan (GXM) is the primary constituent of the capsule and accumulates in the serum and cerebrospinal fluid (CSF) [25,26]. High GXM levels are associated with many immunosuppressive effects, including interference with phagocytosis, antigen presentation, leukocyte migration and proliferation, and specific antibody (Ab) responses, contributing directly to C. neoformans pathogenesis [27]. GXM even enhances HIV replication [28], but despite this knowledge, many specific effects of GXM on C. neoformans brain pathogenesis or inflammatory response remain largely unresolved.
Cryptococcal meningoencephalitis (CME) is uniformly fatal if left untreated, and even with optimal antifungal treatment, mortality ranges from 30-100% [29][30][31]. Non-HIV CME patients display greater brain tissue inflammation and fewer extracellular cryptococci than HIV-associated CME victims [32]. The latter group displays numerous cryptococci in the subarachnoid space post-mortem, associated with raised intracranial pressure (ICP) and hydrocephalus [33], CSF fungal burden, and GXM accumulation, which are all associated with mortality [34]. In addition, blood vessel occlusion and damage are frequently associated with systemic infection, given that individuals with CME develop aneurisms or subarachnoid hemorrhage [35]. In fact, there are instances where CME has been associated with ischemic stroke with vasculitis in HIV + [36] and HIVindividuals [37]. Nevertheless, the mechanisms of CME pathogenesis and the progression of its manifestations are poorly understood.
In this study, we characterized the clinical and pathological manifestations of CNS cryptococcosis in an experimental mouse model of intracerebral (i.c.) infection using stereotaxic surgery. Our results showed that depending on the brain region where C. neoformans infection was found, cerebral cryptococcosis develops with or without brain parenchyma involvement. C. neoformans-infected mouse brains displayed marked similarities with histopathological and clinical manifestations observed in humans, making our model an excellent method to distinguish the direct neurotoxic impact of fungal infection, GXM release, and disease progression. Future research using this mouse model of i.c. fungal infection may help fill in the knowledge gap in our understanding of how cerebral cryptococcosis progresses and may provide novel research avenues in C. neoformans pathogenesis.

i.c. C. neoformans infection in different brain regions of C57BL/6 mice results in hydrocephalus, subarachnoid hemorrhage, and death
We investigated the development of cerebral cryptococcosis after i.c. C. neoformans infection (10 4 fungi of strain H99 in 1 μL of sterile saline) in C57BL/6 mice. We injected the fungus in different regions of the brain including the midline at bregma level, neocortex, striatum, and ventricles (n = 10 mice per region, Fig 1A) using a stereotaxic apparatus (S1 Fig). The bregma is an anatomical landmark in the skull where the sagittal and coronal sutures intersect joining the parietal and frontal bones [coordinates of injection: medial/lateral (X-axis): 0, anterior/posterior (Y-axis): 0, dorsal/ventral (Z-axis): -2.5); Table 1; Fig 1A, upper left panel]. The cerebral cortex is the largest site of neural integration in the CNS playing key roles in attention, perception, awareness, memory, language, and consciousness [coordinates of injection: medial/lateral (X-axis): +/-1.5, anterior/posterior (Y-axis): 0, dorsal/ventral (Z-axis): -1); Table 1; Fig 1A, upper right panel]. The striatum serves as the primary input to the rest of the basal ganglia, receives glutamatergic and dopaminergic inputs from different sources, and is a critical component of the motor and reward systems [coordinates of injection: medial/lateral (X-axis): +/-2, anterior/posterior (Y-axis): 0.2, dorsal/ventral (Z-axis): -3.5); Table 1; Fig 1A,   infected with a single inoculum of 10 4 C. neoformans strain H99 cells in either bregma, cortex, striatum (basal ganglia), or ventricle after a stereotaxic surgery. (B) Survival differences of C57BL/6 mice infected with 10 4 fungi (n = 10 mice per group) in bregma (circle), cortex (square), striatum (triangle), or ventricle (diamond). The survival curves for cortical and ventricular infections are overlapped. Symbols ( � , #, and &) indicate P value significance (P < 0.05) calculated by log rank (Mantel-Cox) analysis. � , #, and & indicate significantly longer survival than in the cortex-, ventricle-, and bregmainfected groups, respectively. (C) Fungal burden in whole brain tissue (numbers of colony forming units (CFU)/ milligram (mg) of tissue) collected from bregma-, cortex-, striatum-, and ventricle-infected mice with 10 4 cryptococci (n = 5 per group) 3-and 7-days post-infection (dpi). (D) Photograph of a mouse infected with C. neoformans H99 in the striatum of the basal ganglia showing signs of hydrocephalus (right) including bulging of the forehead (black arrows). A picture of a healthy uninfected mouse (left) is shown for comparison. These animals were euthanized 20-dpi. (E) Representative brain coronal histology sections of uninfected control and bregma-, cortex-, striatum-, and ventricle-( Fig 1B). The clinical symptoms shown by mice in these groups include respiratory distress, dehydration, severe hunching, dyspnea with considerable abdominal involvement, abnormal neurologic movement with consistent body twitching, and body condition score of 2.5 ( Table 2). All the animals infected in bregma died by 12-dpi. These mice experienced dehydration, hunching, lethargy, body condition score of 2, and a firm cranial deformity of 1.3 mm in diameter (Table 2). Mice infected in the striatum survived the longest, 20-dpi with a median survival of 15-dpi ( Fig 1B). Mice infected in the basal ganglia demonstrated a plethora of clinical signs likely due to their prolonged survival including respiratory distress, mild dehydration, severe hunching, dyspnea, palpitations, locomotor disturbance and rapid movement when alert, recumbency with spastic limb movement (Table 2). Mice infected with C. neoformans in specific brain regions (n = 5 per group) exhibited similar fungal burden 3-dpi (bregma, 3.38 x 10 2 colony forming units (CFU) per gram (g)/tissue; cortex, 3.92 x 10 2 g/tissue; striatum, 4.47 x 10 3 g/tissue; ventricle, 1.43 x 10 3 g/tissue) and 7-(bregma, 3.31 x 10 4 CFU per g/tissue; cortex, 2.39 x 10 4 g/tissue; striatum, 4.29 x 10 4 g/tissue; ventricle, 5.7 x 10 4 g/tissue), regardless of the time of death of all the animals within a specific group (Fig 1C). C57BL/6 rodents infected with C. neoformans in the striatum developed signs of hydrocephalus at 12 to 14 dpi such as bulging of the forehead (black arrows; right panel) and the dome-shaped cranium relative to sterile saline-injected and uninfected control animals (saline-injected/uninfected control; left panel; Table 2; Fig 1D). A representative brain coronal tissue section of saline injected/uninfected mice displayed normal lateral ventricles (black arrows), septum pellucidum (black arrowhead), and parenchyma ( Fig 1E). In contrast, brain tissue from infected groups with C. neoformans evinced obvious signs of hydrocephalus including ventriculomegaly in both lateral and third (red arrows) ventricles, thinning of the cerebral cortex (dashed circles), and widely distributed pale and red microinfarcts throughout the brain parenchyma. Mice infected in bregma, cortex, and striatum also showed cryptococcoma or C. neoformans brain-lesion formation and expansion in tissue due to encephalomalacia or liquefactive necrosis. Gross morphology of representative brains excised from mice infected in bregma (B), cortex (C), striatum (S), and right ventricle (V) upon termination (day 7-dpi for B, C, and V and 20-dpi for S) after perfusion and fixation with 4% paraformaldehyde revealed swollen cerebral infected mice. Uninfected mice exhibit normal brain parenchyma, regular lateral ventricles (black arrows) and an intact septum pellucidum (arrowhead). Dashed black ellipses demonstrate thinning of the cortex. Red arrows indicate ventriculomegaly of the third ventricle. Scale bar: 2 mm. (F) Gross morphology of control and bregma-, cortex-, striatum-, and ventricle-infected brains after perfusion with 4% paraformaldehyde. For E, uninfected (control) and infected mice in bregma, cortex, and ventricle were euthanized 7-dpi. Striatum-infected rodents were sacrificed 20-dpi. (G) Representative periodic acid-schiff (PAS; fungi; red staining; scale bar: 100 μm)-stained sections of the brain are shown and were removed 7-dpi. Lower panel image is a magnification of the smaller box in the upper PAS-stained section to better show the cryptococcal cells (indicated with arrow heads; scale bar: 20 μm). (H) Brain weight determinations from tissue removed 7-dpi. For C and H, bars and error bars denote the means and standard deviations (SDs), respectively. Each symbol represents values for an individual mouse (n = 5 for C and n = 3 for H). Asterisks denote P-value significance ( �� P < 0.01, ��� P < 0.001, and ���� P < 0.0001) calculated using analysis of variance (ANOVA) and adjusted by use of the Tukey's post-hoc analysis. ns denotes not statistically significant comparisons. These experiments were performed twice, similar results were obtained each time, and all the results combined are presented.
https://doi.org/10.1371/journal.pntd.0011068.g001 hemispheres with compressed olfactory bulb, subarachnoid hemorrhage, clotted blood at the superior sagittal sinus and transverse sinus (Fig 1F), in comparison to saline injected/uninfected control (Ctr) mice ( Fig 1F). Periodic acid-schiff (PAS)-stained sections of the excised brains 7-dpi were performed and confirmed the presence of cryptococci (black arrowheads, red staining) in the brain-lesions ( Fig 1G). Finally, weight determinations of brains removed 7-dpi from mice infected with C. neoformans in bregma (P < 0.01), cortex (P < 0.001), striatum (P < 0.001), or right ventricle (P < 0.0001) demonstrated a significant increase in tissue mass relative to saline injected/uninfected control mice ( Fig 1H). There were no differences in brain weight among the infected groups. Our data demonstrate that mice directly infected with C. neoformans in specific brain regions develop signs of hydrocephalus and subarachnoid hemorrhage resulting in mortality.

Characterization of cerebral cryptococcosis in mice infected in specific brain regions
We described the development of cerebral cryptococcosis in coronal brain sections excised from C57BL/6 mice directly infected in bregma 7-dpi. Mice exhibited encapsulated yeast cells lodged in the septum pellucidum inducing encephalomalacia and scattered microinfarcts in superficial (e.g., cerebral cortex and corpus callosum) and deep tissue (e.g., thalamus, hypothalamus, and basal ganglia) (Fig 2A). Higher magnification images showed expanding encephalomalacia or liquefactive necrosis in the brain parenchyma ( Fig 2B) and minimal inflammation ( Fig 2C). GXM-stained immunohistochemistry (IHC) using monoclonal Ab (mAb) 18B7 (Fig 2D), which is specific to the main component of C. neoformans capsular polysaccharide (CPS), revealed wide distribution of GXM especially in the cortical region above of each ventricle, septum pellucidum (encephalomalacia; Fig 2E and 2F), and adjacent blood vessels, particularly in the margins of microinfarcts. We also examined glial responses (microglia and astrocytes) to the fungus at 7-dpi. IHC against the microglia-specific marker ionized calcium binding adaptor molecule 1 (Iba-1, Fig  2G-2I), demonstrated microglial response against the progressive encephalomalacia and cryptococcal growth (Fig 2H). Microglia near fungal cells exhibited a dystrophic or necrotic morphology ( Fig 2I). Similarly, IHC against astrocyte-specific marker glial fibrillary acidic protein (GFAP, Fig 2J-2L) showed an increased number of astrocytes or astrocytosis and fibrillary processes per astrocytes or astrogliosis surrounding the region of cryptococcal-induced encephalomalacia and GXM accumulation (Fig 2K-2L). Astrocytes also appear necrotized and unable to form a barrier between the damaged brain tissue and the remaining tissue ( Fig 2L). These data indicated that mice infected in bregma develop cerebral cryptococcosis characterized by encephalomalacia near the septum pellucidum, GXM accumulation, and glial activation in the marginal zone of the affected tissue.
The cerebral cortex is one of the brain regions affected by CME [32,38]. Thus, we characterized the progression of cerebral cryptococcosis in mice infected with the fungus in the cortex (Fig 3). As shown, a coronal section from a cortically infected mouse brain removed at 7-dpi displayed a large cryptococcoma in the ipsilateral cerebral cortex (Fig 3A; black square), smaller disseminated brain-lesions filled with cryptococci, and signs of ischemia including widespread microinfarcts in deep tissue including the septal nucleus, striatum, and piriform cortex. High magnification images show a cryptococcoma characterized by liquefactive necrosis (Fig 3B), presence of vacuolated or hueco cells, and the remaining tissue in the margins displaying severe degenerative changes with no inflammatory infiltrate ( Fig 3C). The contralateral cerebral cortex (green square in Fig 3A) evinced severe interstitial edema forming status spongiosis in the brain parenchyma ( Fig 3D). The leptomeninges (red square in Fig 3A)  showed edema and phagocytic cell recruitment particularly in the pia matter and arachnoid meninges ( Fig 3E). Subarachnoid hemorrhage was observed (blue square in Fig 3A) and extravasated blood accumulated (white arrows) between pia matter and arachnoid meninges ( Fig 3F). IHC demonstrated ubiquitous ipsilateral and contralateral distribution of GXM with more prominence in the margins of infarction (Fig 3G-3I). Iba-1 staining displayed activated microglia with many and thick ramifications in response to tissue damage due to cryptococcal colonization (Fig 3J-3L). Dystrophic or necrotic microglia can be observed in tissue surrounding the region of expanded encephalomalacia ( Fig 3L). Substantial astrocytosis and astrogliosis was shown surrounding the cryptococcal brain lesion and clumping in cortical tissue (Fig 3M-3O). Our findings indicate that mice i.c. infected in the cortex developed local and scattered cryptococcomas characterized by encephalomalacia, edema, ischemic signs including blood vessel infarcts, meningeal bleeding, significant CPS accumulation, and glial morphological changes.
In post-mortem brain tissue from CME patients, cryptococci are usually distributed into the corpus striatum of the basal ganglia, where dopaminergic neurons are abundant [32]. Given the importance of this region in human disease, we documented the advancement of the cryptococcal infection in the striatum of C57BL/6 mice 20-dpi. A representative brain coronal section at midpoint of the infundibulum displayed aggressive cryptococcoma expansion from the striatum to include the thalamus, hypothalamus, habenula, the microinfarction in contralateral thalamus and hippocampus and frontoparietal cortex ( Fig 4A). The area of ischemic or liquefactive necrosis displayed hemorrhage (Fig 4B-4C), vacuolated macrophages or hueco cells (black arrowheads) and foreign body giant cells (red arrow; Fig 4D). We observed leptomeningitis with infiltration of activated epithelioid or macrophage-like cells (red square in Fig 4A; black arrow; Fig 4E). Perivascular edema and ischemic or angular neurons (white arrows) developed in the contralateral frontoparietal and visual cortex suggesting obstruction of venous drainage system (blue square in Fig 4A and 4F). GXM staining the margins of the cryptococcoma, extending to the frontoparietal cortex of the ipsilateral and contralateral sides and the thalamus of the contralateral side (Fig 4G-4I). High magnification images show GXM intensely distributed around the margins of the cryptococcoma and concentrated in the wall of cerebral blood capillaries surrounding the area of encephalomalacia (Fig 4H-4I). Microglia are observed surrounding the cryptococcoma (Fig 4J). These brain surveillance resident cells are shown activated inside the cryptococcoma ( Fig 4K) and dystrophic surrounding the liquefactive tissue ( Fig  4L). Significant and extensive astrocytosis is shown in the wall of the cryptococcoma that extends to the hippocampus (Fig 4M). The presence of astrogliosis including astrocytic clumping surrounding the margins of the cryptococcoma is evident (Fig 4N-4O). These results show that direct infection of the basal ganglia is characterized by significant expansion of the cyptococcoma that consists of yeast cells, necrotic tissue with hemorrhage, hueco cells, and foreign body giant cells. Furthermore, the development of meningitis and edema with ischemic neurons can be appreciated especially in brain regions distant from the initial inoculum.
C. neoformans infects the choroid plexus and the ventricles, causing ventriculitis and/or hydrocephalus. This condition is associated with increased ICP [39,40]. Lack of CSF reabsorption in the arachnoid granulations due to the accumulation of cryptococcal polysaccharide and/or yeast cells has been related to increased ICP [41,42]. The presence of CPS in the CSF is used for diagnosing CME in patients, highlighting the importance of understanding how the ventricle environment is affected during C. neoformans infection. Therefore, we also studied the progression of C. neoformans infection in the brain right ventricle of C57BL/6 mice after tissue removal 7-dpi. A coronal section of brain tissue at midpoint of the infundibulum shows ventriculomegaly, infarction in the hippocampus, and substantial subarachnoid hemorrhage (black arrow; Fig 5A). Minimal cryptococcoma formation was observed in mice infected in the right ventricle. The area of infarction in the hippocampus exhibited liquefactive necrosis ( Fig 5B). A high magnification image demonstrates minimal inflammation and the presence of a megakaryocyte (blue arrow) that is indicative of thrombosis ( Fig 5C). Ventricle-infected tissue demonstrated extensive distribution of GXM in regions such as the cerebral cortex, meninges, basal ganglia, and corpus callosum (Fig 5D). GXM was concentrated in the frontoparietal cortex and the wall of large blood vessels and capillaries (Fig 5E and 5F). Microglia was mostly concentrated in the hippocampus around the infarction area (Fig 5G-5I). In contrast, astrocytes were distributed all over the brain tissue with astrogliosis mostly observed near blood vessels (Fig 5J-5L). Our observations in ventricle-infected mice demonstrate minimal cryptococcoma formation, minimal inflammation, localized infiltration of microglia, and significant accumulation of GXM especially in blood vessels, which are surrounded by reactive astrocytes.

Relationship between C. neoformans GXM localization and glial cell distribution in the mouse brain
We analyzed the intensity correlation of either Iba-1 and GFAP markers relative to C. neoformans GXM in coronal tissue sections excised from mice infected in the bregma, cortex, striatum, or ventricles using the NIH Image J software (Fig 6). Mice infected in bregma showed positive correlations between Iba-1 (R 2 = 0.3533; P = 0.0152; Fig 6A) and GFAP (R 2 = 0.2942; P = 0.0299; Fig 6B) stained cells and GXM distribution. Iba-1 and GXM (R 2 = 0.1269; P = 0.1468; Fig 6C) staining did not correlate in mice infected in the cerebral cortex, although there was a correlation between the GFAP and GXM (R 2 = 0.2472; P = 0.0358; Fig 6D) intensity. In addition, mice infected in the striatum evinced positive correlations between the intensity of Iba-1 (R 2 = 0.5915; P = 0.0035; Fig 6E) and GFAP (R 2 = 0.4958; P = 0.0106; Fig 6F) staining and the intensity of GXM in tissue. There was no correlation between Iba-1 (R 2 = 0.0009; P = 0.8915; Fig 6G) and GFAP (R 2 = 0.0671; P = 0.2215; Fig 6H) intensity and GXM in ventricle-infected mice. These analyses demonstrate that glial cell responses to GXM accumulation are different depending on the localization of the C. neoformans infection in the brain.

Mortality is associated with GXM localization and hemorrhage in the subarachnoid space
Approximately 30% of clinical cases have reported infarcts and blood vessel damage in patients with severe CME [43]. Histopathological examinations of mice infected i.c. with C. neoformans at the time of death demonstrated that mortality was associated to hemorrhage in the brain subarachnoid space and GXM accumulation in this tissue area (Fig 7). Brain coronal tissue section at midpoint of the infundibulum from an uninfected and healthy mouse showed normal anatomy and normal leptomeninges covering the retrosplenial and frontoparietal cortex (Fig 7A and 7B). Mice infected in bregma (Fig 7C), and brain tissue removed 7-dpi displayed notable subarachnoid hemorrhage (black arrows; Fig 7D) and moderate GXM deposition in the glia limitans and meninges (red arrows; Fig 7E). Cortical tissue excised 7-dpi demonstrated marked hemorrhage in the subarachnoid space (Fig 7F-7G), bulging of the superior sagittal sinus (black arrow; Fig 7F-7G), and extensive GXM distribution in the leptomeninges especially near the bleeding (red arrows; Fig 7H). Moreover, mice infected with the fungus in the striatum (Fig 7I-7K) or ventricles (Fig 7L-7N) also showed considerable subarachnoid space and superior sagittal sinus bleeding (black arrows; Fig 7J and 7M). GXM was profusely dispersed in the leptomeninges (red arrows; Fig 7K and 7N). Our examinations suggest that mice mortality may be related to extensive C. neoformans CPS localization near the subarachnoid space blood vessel infarcts and hemorrhage.

C. neoformans cortical infection results in apoptosis of neurons in the hippocampus
We further investigated the sequalae of cerebral cortical C. neoformans infection in C57BL/6 mice. Histopathological examination of the hippocampus, one of the main brain regions associated with memory and cognitive function, after fungal infection of the cerebral cortex showed severe edema and passive hyperemia likely due to obstruction of cerebral blood flow ( Fig 8A). We also observed signs of neurodegeneration in the dentate gyrus, appearing as cytoplasmic vacuolation. High power magnification (white rectangle box) also indicated apoptotic pyramidal neurons (red arrows) distributed across the CA3 region (Fig 8B), suggesting the possibility that cerebral cryptococcosis, particularly cortical colonization, may cause altered behavior in mice or altered mental status in patients with CME.

C. neoformans i.c. infection causes microinfarctions and tissue bleeding
We documented the effects of C. neoformans i.c. infection in brain tissue (Fig 9). Hemorrhagic microinfarctions were observed in the frontal cortex (rectangle; Fig 9A and 9B) and increased infiltration of gitter cells (blue arrows; Fig 9C), which are enlarged phagocytic cells of microglial origin having the cytoplasm distended with lipid granules. Prussian blue staining confirmed hemosiderosis or excessive accumulation of iron deposits that arises from hemorrhage (Fig 9C; inset). Hemorrhagic microinfarctions were visualized near to cryptococcomas in the corpus callosum (Fig 9D). We observed in this region, leukomalacia or necrosis of the white

PLOS NEGLECTED TROPICAL DISEASES
matter, degeneration and demyelination of the nerve fibers, and recruitment of gitter cells ( Fig  9E) in the presence of erythrocytes (black arrows; Fig 9F). Hemorrhage in the margins of a cryptococcoma (rectangle box; Fig 9G) shows extensive aggregation of gitter cells around the bleeding area and cryptococcal brain lesion (Fig 9H). High power magnification confirmed significant gitter cell recruitment (blue arrows) into the hemorrhage region (Fig 9I). i.c. C. neoformans infection of the striatum results in extensive encephalomalacia, multiple hemorrhage foci (Fig 9J), and considerable aggregation of polymorphonuclear cells around the blood vessels and in the lesion (Fig 9K-9L). A coronal section of mice infected with C. neoformans show multiple cryptococcomas scattered in the posterior limb of internal capsule, the putamen

Characterization of Cryptococcus neoformans intracerebral infection
(black rectangle box), nucleus accumbens, and septal nucleus (Fig 9M). Microthrombus in the wall of a blood capillary and well circumscribed infarcted tissue from ischemia and liquefactive necrosis (red arrows; Fig 9N). These images show that cerebral cryptococcosis after i.c. infection causes thrombosis of blood capillaries in the subarachnoid spaces resulting in ischemia, microinfarctions, and hemorrhage in regions of the CNS where there is fungal colonization and adjacent GXM accumulation.

C. neoformans GXM extensively deposits in the vasculature
C. neoformans GXM is extensively released and accumulates in the CSF and brain tissue during infection. Recently, it was shown in zebra fish that blood vessel occlusion by this encapsulated fungus is a mechanism for hemorrhagic dissemination of infection [44]. Given that GXM is associated with excessive bleeding in the subarachnoid space and increased mortality in mice i.c. infected, we examined the localization of CPS in brain tissue with emphasis in the vasculature. GXM IHC with mAb 18B7 of a coronal brain section of a mouse infected in the right ventricle with C. neoformans showed dispersed polysaccharide throughout the tissue and in the wall of blood vessels in the cerebral cortex, hippocampus, habenula, thalamus and hypothalamus ( Fig 10A). Considerable GXM deposition was visualized in the wall of blood capillaries (red arrow) and large vessels (black rectangle box; Fig 10B) in the thalamus. Higher magnification image demonstrated significant GXM deposition (brown stain) in the walls of a medium size blood vessel and several surrounding capillaries (Fig 10C). Furthermore, GXM accumulated mainly in the wall of large blood vessels in the subarachnoid space (red arrows; Fig 10D). Our findings that the fungal CPS localizes in the vasculature provides supportive evidence that GXM and cryptococci may accumulate and obstruct blood vessels in the CNS resulting in microinfarctions, hemorrhage, and exacerbating death.

Mice infected with C. neoformans i.c. develop ischemic necrosis
Histopathological examinations of the frontal cortex from C57BL/6 mice infected with C. neoformans demonstrate noticeable ischemic necrosis of the gray matter (black box rectangle; Fig 11A). Ischemic necrosis is consistent with the presence of clear spaces of liquefactive tissue where the damaged neurons and tissue are phagocytosed and then removed. High power magnification shows infiltration and accumulation of gitter cells (blue arrows) with foamy cytoplasm that results from engulfing necrotic tissue (Fig 11B). In addition, microinfarctions were observed in cortical tissue, characterized by the presence of pleiomorphic or distorted erythrocytes (red arrows; Fig  11C). A C. neoformans lesion or cryptococcoma in the striatum of mice displayed a foreign body giant cells (black rectangle) surrounded by fungal cells and numerous phagocytic cells with vacuolated cytoplasm (Fig 11D-11F). Damage of cerebral capillaries (encircled red areas) and microthrombus in the striatum of infected mice are shown with presence of a few Langhans giant cells (red arrows) and little macrophagic infiltration indicating poor granuloma formation (Fig 11G). Also, coagulative necrosis is observed in neurons with pyknotic nuclei and increased cytoplasmic acidophilia. There is noticeable accumulation of cryptococci in liquefactive tissue around the border of the brain lesion likely waiting for more liquefaction of tissue to proliferate further. Our results show that C. neoformans cerebral colonization causes extensive vascular damage and ischemic necrosis, two observations associated with mice mortality.

Microglia inside of a cryptococcoma becomes activated while combatting the infection
We documented the microglial responses to C. neoformans i.c. infection in the striatum (Fig 12). Histopathology showed phagocytic cells engulfing cryptococcal cells and necrotic

PLOS NEGLECTED TROPICAL DISEASES
Characterization of Cryptococcus neoformans intracerebral infection tissue in the cryptococcoma (Fig 12A). The presence of free yeast cells in the brain lesion and phagocyte recruitment to the damage tissue or the area of encephalomalacia is evident (Fig 12A). IHC for Iba-1 in the cryptococcoma displays microglia or hueco cells with vacuolated cytoplasm (red arrows) responding to the progressive encephalomalacia (Fig 12B). We identified engulfed cryptococci by several microglia in the cryptococcoma lesion. These images show that microglia or hueco cells are one of the hallmarks of a cryptococcoma. They activate and try to fight the infection by this encapsulated fungus while removing necrotic or damaged tissue. Discussion CME is the most important manifestation of C. neoformans infection. Most of our understanding of neurocryptococcosis comes from post-mortem observations in humans [32,38]. One of the main roadblocks to study CME is the availability of well-characterized or reliable animal models where the disease progression and dynamics can be systematically studied, and the pathological manifestations show similarities to those presented in humans. Although not a natural route of C. neoformans infection, we used the i.c. model to study the development of CME and the cellular/physiological changes/responses in specific regions of the brain with a controlled inoculum (10 4 fungi) and a suitable volume (1 μL). Stereotaxic injections, although invasive, are extensively used in neuroscience research to gain knowledge of CNS infectious disease processes. In fact, tissue damage/response from the injection can be easily controlled by using saline-injected and naïve animals. In addition, i.c. infection models have been reported in multiple diseases including cryptococcosis [45][46][47][48], although not clearly described in studies with C. neoformans. An advantage of the i.c. C. neoformans model of infection over the intranasal or intratracheal (pulmonary) and intravenous (systemic) models is that the fungal brain dissemination number and the brain region location of infection can be controlled. This is important because the effect and progression of the cryptococcal infection in specific regions of the brain can be characterized. The impact of CME on behavior alteration including motor and cognition can be simply examined. Furthermore, the i.c. model can be combined with CD4-specific mAb administration, which depletes 95% of CD4 + T cells, mimicking the reduction of these lymphocytes in HIV infection [49,50]. In contrast, the need of mouse-specialized surgical expertise, basic to intermediate knowledge of neuroscience, and the availability of specific equipment such as a stereotaxic apparatus are perhaps the main disadvantages.
Cerebral cryptococcosis in humans is characterized by meningoencephalitis with (cryptococcoma) and without (no cryptococcoma) involvement of the brain parenchyma, especially in immunosuppressed and immunocompetent individuals, respectively [32]. There are multiple mechanisms proposed that can lead to meningitis including rupture of cryptococcomas and spread of infection from the perivascular space to the CNS [10,[51][52][53]. Mice infected in the cortex and ventricles with C. neoformans showed early mortality, although they exhibited different disease development. Rodents infected in cortical tissue demonstrated cryptococcoma formation, hydrocephalus, edema, and subarachnoid hemorrhage while microinfarctions were minimal. Meningitis in these mice was prominent considering that the meninges directly cover the cortex, and cryptococci and GXM easily reach these membranes through circulation. Similarly, mice infected in the striatum displayed comparable pathology to those infected in the cortex, with the exception that they began dying approximately a week later despite development of microinfarctions. The last mouse died 12 days after those infected in cortical tissue. It is plausible that infection in the striatum, being in a deep region of the brain, takes longer for the fungus to first colonize and form a cryptococcoma, release sufficient GXM that impairs the migration and effector functions of microglia and other inflammatory cells, and cause tissue necrosis. It is likely that free yeast cells or yeast cells engulfed by microglia escape the biofilm-like lesion and enter the capillary network of the CNS, reaching the subarachnoid space and resulting in meningitis and hemorrhage. In this regard, we found similar fungal loads and brain weights in all the routes of infection, validating this possibility. In contrast, intrathecal infection in the ventricles evinced comparable pathologies as those documented in mice infected in the cortex or striatum without brain parenchyma involvement or no cryptococcoma formation. When C. neoformans is injected into the ventricles and directly delivered into the CSF, cryptococci reach the meninges and subarachnoid space, rapidly killing the mice (at~8-dpi). Although there is no cryptococcoma formation like in humans [54], GXM diffuses profusely throughout the tissue, showing affinity for blood vessels where the CPS accumulates [55]. It would be interesting to infect the mice intrathecally with a lower inoculum (10 3 cryptococci) to determine if survival is prolonged to study the infection longer and still have no brain parenchyma involvement. It is apparent from our studies that the region of cryptococcal CNS access as well as the colonization process are important in the progression of CME including the live or death outcomes in patients. Our studies demonstrate, both, cortex and striatum, routes of cryptococcal i.c. infection represent acceptable models to study CME with brain tissue involvement whereas ventricular infection is appropriate to study cerebral cryptococcosis limited to the subarachnoid space.
Individuals with HIV + and CME show little inflammation in the CNS due to their reduced number of CD4 + T cells [32]. Our findings demonstrated that regardless of the location of the brain infection, cryptococcal infection shows minimal inflammation independent of whether there is tissue or just subarachnoid space involvement. Given that C. neoformans extensively releases its GXM that accumulates expansively around cryptococcomas or throughout brain tissue, this CPS inhibits inflammatory responses including recruitment and effector functions such as phagocytosis particularly by microglia [56], the resident cells of immunity in the CNS. GXM inhibits neutrophil migration to microglial-produced IL-8 [57] and the rolling and attachment of these leukocytes to endothelial cells by interfering with their adhesion molecules [58,59]. C. neoformans' neurotropism may be related to the inability of the host in mounting an effective neuroimmune response against the fungus, particularly in immunosuppressed individuals. Lee and colleagues previously suggested that the impaired immunity in patients with AIDS promotes the accumulation of mostly extracellular C. neoformans within the brain [32], and that deficient macrophage/microglial effector function may be responsible for the altered pathology. Our histological data show ischemia in tissue with cryptococcomas that was characterized by the recruitment and accumulation of gitter cells with foamy cytoplasms that result from engulfing necrotic tissue and considerable vascular damage. Ischemic stroke is a recognized complication of CME in the acute phase and is thought to be mediated by an infectious vasculitis [60]. C. neoformans can also induce extensive fibrosis of the subarachnoid space, which may compress small veins mechanically inducing venule congestion and massive cerebral infarction [61]. We documented proliferation and accumulation of cryptococci in liquefactive tissue around the border of the brain lesion and considerable deposition of GXM in blood vessels, which may cause compression and occlusion resulting in thrombosis and microinfarctions. In this regard, a recent study on zebra fish demonstrated that cryptococci become trapped in blood vessels, proliferate, cause vascular damage, cortical microinfarctions, and hemorrhage during CME [44]. Structural studies on blood vessels and endothelial cells suggest that C. neoformans and possibly GXM cause alterations of tight junction and adhesion proteins impairing cell-cell interactions at the blood-brain barrier interface and vasodilation, which affect the vascular tension and facilitates fungal dissemination [44,62].
The role of glia in CME is poorly understood. Post-mortem tissue from humans have shown the association of glial cells and GXM deposition [55]. We established that glial cell responses to the accumulation of CPS is dependent on the CNS region infected by C. neoformans. High deposition of GXM in bregma and the striatum correlated with the presence of both microglia and astrocytes. GXM accumulation in cortical tissue correlated with astrocyte presence but not with microglia. The deposition of CPS in mice infected in the ventricles did not correlate with astrocytes nor microglia. It is possible that the prolonged survival documented in mice infected in bregma and the striatum is associated with the astrocyte/microglia responses and their communication with each other or with other cells including those recruited from the periphery. For example, CPS immunoreactivity has been documented in activated microglia interacting with cryptococcal lesions or associated with destructive parenchymal lesions such as necrotizing tissue or infarcts [55]. Similarly, GXM has been identified inside of microglia and bound to the cell body of reactive astrocytes [55]. However, the mechanisms by which the CPS impact glial cell functions are not well-understood. Thus, further investigations are needed to elucidate the importance of a concerted astrocyte/microglia response against C. neoformans in the CNS.
In conclusion, we developed and systematically described a murine model to study the dynamics of cerebral cryptococcosis. Our model recapitulates many of the human pathological manifestations depending on the brain region where the fungus is delivered, including the presence of foreign body giant cells, vacuolated microglia or hueco cells, granulomatous meningitis, neuronal necrosis, and accumulation of the CPS in blood vessels. In addition, the pathogenesis of cerebral cryptococcosis progresses differently depending on the region of infection, making these models suitable to study CME with (e.g., cryptococcoma or cryptococcal brain lesions) or without involvement of the brain parenchyma that result in meningitis, edema, hydrocephalus, and subarachnoid hemorrhage leading to death. Investigations adopting this mouse model of cryptococcal infection may assist us in elucidating critical questions associated with this medically important mycosis, as well as the development of new therapeutics and treatments to combat the CNS disease, which is responsible for significant morbidity and mortality.

Ethics statement
All animal studies were conducted according to the experimental practices and standards approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Florida (Protocol #: 202011067). The IACUC at the University of Florida approved this study.

C. neoformans
C. neoformans strain H99 (serotype A) was isolated and kindly provided by John Perfect at Duke University. Yeasts were grown in Sabouraud dextrose broth (pH 5.6; BD Difco) for 24 h at 30˚C in an orbital shaker (Thermo Fisher) set at 150 rpm (to early stationary phase).

i.c. infection with C. neoformans, animal care, and survivability studies
The equipment setup for mouse stereotaxic surgeries and i.c. infections are illustrated in the S1 Fig. C57BL/6 female mice (6-8 weeks old; Envigo) were anesthetized using isoflurane (3-5% for induction and 1-2% maintenance; model: VetFlo Vaporizer Single Channel Anesthesia System, Kent Scientific), placed in prone position over a heating pad (model: RightTemp Jr., Kent Scientific), and prepped using aseptic techniques. A local anesthetic, bupivacaine or ropivacaine (0.05%), was administered subcutaneously in the incision. The fur on the skull was carefully shaved off and the animal was placed in a stereotaxic apparatus (model: 940; Kopf Instruments). The mouse's head was secured using the instrument's ear bars, incisor bar, and nose clamp as is customary for standard stereotaxic procedures. The skin on the shaved head was prepped with 3 alternating, outward spiral applications of 2-4% chlorhexidine scrub and saline prior to incision, and a midline 5-mm incision was made using a scalpel under the guidance of a dissecting microscope. The periosteum was reflected open using small surgical clips. Bregma and lambda skull suture junctions were identified using a stereotaxic brain atlas (e.g., The Allen Mouse Brain Atlas; https://mouse.brain-map.org/static/atlas). Using a small handheld drill (model: Ideal microdrill; Braintree Scientific), the skull was thinned until the underlying dura mater was visible and a micropipette was brought to the correct stereotaxic position and lowered until it touched the exposed dura. The craniotomy was around 1 mm in diameter. The cut tip of the micropipette has a sharp edge, which will allow easy penetration of the dura of mice. Based on previous studies using the Cn H99 strain [63,64], a 1-μL suspension containing 10 4 cryptococci in sterile saline was injected into the brain (e.g., bregma, cortex, striatum, or ventricle; stereotaxic coordinates are shown in Table 1) using a 26 G Hamilton syringe connected to a pump (model: UltraMicroPump3; World Precision Instruments). We injected the fungal inoculum in a 1-μL volume to avoid tissue damage or diffusion of the cryptococci to other regions of the brain. The skin incision on the dorsal head was closed with sterile nylon suture and 2-4% topical chlorhexidine solution was be applied over the closed incision. After the surgery, mice were placed on a clean recovery cage and monitored for survivability. The survival end points were inactivity, tachypnea, or loss of � 25% of body weight from baseline weight. We monitored the mice twice daily for clinical signs, dehydration, and weight loss ( Table 2). Animals showing signs of dehydration or that lost more than 10% weight received supportive care such as 1 mL of parenteral fluid supplementation (saline) and moist chow on the cage floor was provided. In separate infections, brain tissues were excised for processing for determination of CFU numbers and histopathological studies.

Photographs
The surgical procedure, i.c. injections, physical changes (e.g., hydrocephalus), and excised brains were photographed with a digital camera and the pictures were transferred to a secured computer. We followed the recommendations from the University of Florida Animal Care Services Policy.

CFU determinations
Brains were excised from euthanized mice and weighed 3-and 7-dpi. The brain tissue was homogenized in 5 mL of sterile phosphate buffered saline, serially diluted, a 100 μL suspension was plated on Sabouraud dextrose agar (BD Difco) and incubated at 30˚C for 48 h. Quantification of viable yeast cells from infected animals were determined by CFU counting.

Histological examinations
Brain tissue was excised from euthanized mice 7-and 20-dpi and fixed in 4% paraformaldehyde (Sigma) for 24 h. Tissues were managed, embedded in paraffin, and 4 μm coronal sections were cut and fixed to glass slides and then subjected either to hematoxylin and eosin (H&E) stain, to PAS stain, or to Prussian blue stain to examine host tissue, fungal morphology, or hemosiderosis, respectively. The tissues were then stained for GXM, microglia, and astrocytes using a mAb 18B7 (dilution: 1:1,000; anti-cryptococcal GXM IgG 1 generated and generously provided by Arturo Casadevall at Johns Hopkins Bloomberg School of Public Health), Iba-1 (dilution: 1:1,000; FujiFilm Wako Chemicals USA Corp.), and GFAP (dilution: 1:1,000; DAKO), respectively, conjugated to horseradish peroxidase. For histopathological evaluations, whole slides scanning was performed using the Leica Aperio Scanscope CS and analyzed using The Image-Scope software (version 12.4.3). Moreover, each slide was carefully examined with a Leica DMi8 inverted microscope, and photographed with a Leica DFC7000 T digital camera using the Leica software platform LAS X. Each image was blindly analyzed by three independent investigators. The pathological similarities observed in mice and humans with cerebral cryptococcosis were reported (Table 3) by a veterinary pathologist (Mohamed F. Hamed) and physician specialized in infectious diseases (Mircea R. Mihu) who independently analyzed the data.

Analysis of GXM and glia colocalization in brain tissue
Images of the brain tissue stained with GXM-, Iba-1 (microglia)-, and GFAP (astrocytes)-specific mAbs were analyzed to correlate GXM to glia colocalization by comparing differences in stain intensity in select regions of mice infected with C. neoformans in bregma, cortex, striatum, and ventricle using the NIH ImageJ software (version 1.46r). Each image was sectioned by a 6 x 8 mm square grid, and regions of each tissue sample were selected for analysis of unweighted light intensity using the ImageJ histogram function. The mean light intensities and standard deviations were recorded to identify and compare regions of varying stain intensity between GXM and the microglia or astrocyte stains. Following acquisition of light intensity data for each corresponding section of the stained samples, the data was plotted in Prism 9.4 (GraphPad software). The results were then compared and tested for statistical significance using correlation analysis and simple linear regression.

Statistical analysis
All data were subjected to statistical analysis using Prism 9.4 (GraphPad). Differences in survival rates were analyzed by the log-rank test (Mantel-Cox). P values for multiple comparisons were calculated by analysis of variance (ANOVA) and were adjusted by use of the Tukey's post-hoc analysis. P values for individual comparisons were calculated using student's or multiple t-test analyses. Concentration-response curve for vascular reactivity was log-transformed, normalized to percent maximal response, and fitted using a nonlinear regression. P values of <0.05 were considered significant.
Supporting information S1 Fig. Stereotaxic apparatus and equipment setup used for i.c. C. neoformans infections in different regions of the brain of C57BL/6 mice. (A) Stereotaxic instrument uses a threedimensional coordinate system (e.g., The Allen Mouse Brain Atlas; https://mouse.brain-map. org/static/atlas) to locate specific targets inside the mouse brain. (B) Ultra micro pump holds microinjection syringes to deliver picoliter to milliliter volumes and mounts directly on the stereotaxic frame. (C) Touch screen controller dispenses specific inoculum volume in a microinjection syringe into determined region of the mouse brain.